@aipack-ai/eval 1.1.6 → 1.1.8
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/index.d.ts +6 -1
- package/dist/index.js +5 -1
- package/dist/index.js.map +1 -1
- package/dist/src/cli.d.ts +9 -3
- package/dist/src/cli.js +361 -26
- package/dist/src/cli.js.map +1 -1
- package/dist/src/core/compare.d.ts +28 -0
- package/dist/src/core/compare.js +101 -0
- package/dist/src/core/compare.js.map +1 -0
- package/dist/src/core/embedding.d.ts +33 -0
- package/dist/src/core/embedding.js +68 -0
- package/dist/src/core/embedding.js.map +1 -0
- package/dist/src/core/history.d.ts +39 -0
- package/dist/src/core/history.js +134 -0
- package/dist/src/core/history.js.map +1 -0
- package/dist/src/core/import.d.ts +30 -0
- package/dist/src/core/import.js +111 -0
- package/dist/src/core/import.js.map +1 -0
- package/dist/src/core/judge-llm.d.ts +13 -0
- package/dist/src/core/judge-llm.js +47 -0
- package/dist/src/core/judge-llm.js.map +1 -0
- package/dist/src/core/live.d.ts +111 -0
- package/dist/src/core/live.js +187 -0
- package/dist/src/core/live.js.map +1 -0
- package/dist/src/core/report.d.ts +9 -2
- package/dist/src/core/report.js +30 -5
- package/dist/src/core/report.js.map +1 -1
- package/dist/src/core/runner.d.ts +10 -5
- package/dist/src/core/runner.js +174 -41
- package/dist/src/core/runner.js.map +1 -1
- package/dist/src/core/scorer/index.d.ts +20 -2
- package/dist/src/core/scorer/index.js +36 -9
- package/dist/src/core/scorer/index.js.map +1 -1
- package/dist/src/core/scorer/llm-judge.d.ts +43 -0
- package/dist/src/core/scorer/llm-judge.js +111 -0
- package/dist/src/core/scorer/llm-judge.js.map +1 -0
- package/dist/src/core/scorer/semantic.d.ts +27 -0
- package/dist/src/core/scorer/semantic.js +74 -0
- package/dist/src/core/scorer/semantic.js.map +1 -0
- package/dist/src/core/types.d.ts +65 -4
- package/dist/src/core/validate.js +9 -3
- package/dist/src/core/validate.js.map +1 -1
- package/package.json +4 -3
package/dist/index.d.ts
CHANGED
|
@@ -20,5 +20,10 @@ export type { CaseScore } from './src/core/scorer';
|
|
|
20
20
|
export { validateEvalCase } from './src/core/validate';
|
|
21
21
|
export { loadCases, defaultCasesDir, packageRoot, } from './src/core/loader';
|
|
22
22
|
export type { LoadCasesResult } from './src/core/loader';
|
|
23
|
-
export { runEval } from './src/core/runner';
|
|
23
|
+
export { runEval, skipReason } from './src/core/runner';
|
|
24
24
|
export { renderMarkdown, writeReport, writeBaseline, readBaseline, reportToBaseline, compareBaseline, finalizeReport, } from './src/core/report';
|
|
25
|
+
export type { RenderOptions } from './src/core/report';
|
|
26
|
+
export { LIVE_ENV_KEYS, parseModelSpec, resolveLiveSpec, createLiveLlm, resolveLiveLlm, describeLiveLlm, assertLiveReady, } from './src/core/live';
|
|
27
|
+
export type { LiveLlmSpec, LiveLlm, ResolveLiveOptions } from './src/core/live';
|
|
28
|
+
export { appendHistory, readHistory, renderHistoryTrend, reportToHistoryEntry, sparkline, } from './src/core/history';
|
|
29
|
+
export type { HistoryEntry } from './src/core/history';
|
package/dist/index.js
CHANGED
|
@@ -21,6 +21,10 @@ export { expectedToScorers, resolveScorers, scoreTrace, RULE_SCORER_TYPES, runRu
|
|
|
21
21
|
export { validateEvalCase } from './src/core/validate.js';
|
|
22
22
|
export { loadCases, defaultCasesDir, packageRoot, } from './src/core/loader.js';
|
|
23
23
|
// ─── Runner 与报告 ────────────────────────────────────────────────
|
|
24
|
-
export { runEval } from './src/core/runner.js';
|
|
24
|
+
export { runEval, skipReason } from './src/core/runner.js';
|
|
25
25
|
export { renderMarkdown, writeReport, writeBaseline, readBaseline, reportToBaseline, compareBaseline, finalizeReport, } from './src/core/report.js';
|
|
26
|
+
// ─── live 模式(真实 LLM,M3)─────────────────────────────────────
|
|
27
|
+
export { LIVE_ENV_KEYS, parseModelSpec, resolveLiveSpec, createLiveLlm, resolveLiveLlm, describeLiveLlm, assertLiveReady, } from './src/core/live.js';
|
|
28
|
+
// ─── 历史趋势(日环比可视化,M3)───────────────────────────────────
|
|
29
|
+
export { appendHistory, readHistory, renderHistoryTrend, reportToHistoryEntry, sparkline, } from './src/core/history.js';
|
|
26
30
|
//# sourceMappingURL=index.js.map
|
package/dist/index.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AA0BH,kEAAkE;AAClE,OAAO,EAAE,kBAAkB,EAAE,MAAM,wBAAwB,CAAC;AAC5D,OAAO,EACL,eAAe,EACf,mBAAmB,GACpB,MAAM,uBAAuB,CAAC;AAG/B,iEAAiE;AACjE,OAAO,EAAE,iBAAiB,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAC;AAEtE,kEAAkE;AAClE,OAAO,EACL,iBAAiB,EACjB,cAAc,EACd,UAAU,EACV,iBAAiB,EACjB,aAAa,EACb,YAAY,GACb,MAAM,mBAAmB,CAAC;AAG3B,iEAAiE;AACjE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qBAAqB,CAAC;AACvD,OAAO,EACL,SAAS,EACT,eAAe,EACf,WAAW,GACZ,MAAM,mBAAmB,CAAC;AAG3B,kEAAkE;AAClE,OAAO,EAAE,OAAO,EAAE,MAAM,mBAAmB,CAAC;
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../index.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;GAWG;AA0BH,kEAAkE;AAClE,OAAO,EAAE,kBAAkB,EAAE,MAAM,wBAAwB,CAAC;AAC5D,OAAO,EACL,eAAe,EACf,mBAAmB,GACpB,MAAM,uBAAuB,CAAC;AAG/B,iEAAiE;AACjE,OAAO,EAAE,iBAAiB,EAAE,UAAU,EAAE,MAAM,uBAAuB,CAAC;AAEtE,kEAAkE;AAClE,OAAO,EACL,iBAAiB,EACjB,cAAc,EACd,UAAU,EACV,iBAAiB,EACjB,aAAa,EACb,YAAY,GACb,MAAM,mBAAmB,CAAC;AAG3B,iEAAiE;AACjE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qBAAqB,CAAC;AACvD,OAAO,EACL,SAAS,EACT,eAAe,EACf,WAAW,GACZ,MAAM,mBAAmB,CAAC;AAG3B,kEAAkE;AAClE,OAAO,EAAE,OAAO,EAAE,UAAU,EAAE,MAAM,mBAAmB,CAAC;AACxD,OAAO,EACL,cAAc,EACd,WAAW,EACX,aAAa,EACb,YAAY,EACZ,gBAAgB,EAChB,eAAe,EACf,cAAc,GACf,MAAM,mBAAmB,CAAC;AAG3B,8DAA8D;AAC9D,OAAO,EACL,aAAa,EACb,cAAc,EACd,eAAe,EACf,aAAa,EACb,cAAc,EACd,eAAe,EACf,eAAe,GAChB,MAAM,iBAAiB,CAAC;AAGzB,yDAAyD;AACzD,OAAO,EACL,aAAa,EACb,WAAW,EACX,kBAAkB,EAClB,oBAAoB,EACpB,SAAS,GACV,MAAM,oBAAoB,CAAC"}
|
package/dist/src/cli.d.ts
CHANGED
|
@@ -3,9 +3,15 @@
|
|
|
3
3
|
* packages/eval/src/cli.ts - aipack-eval 命令行入口
|
|
4
4
|
*
|
|
5
5
|
* 用法:
|
|
6
|
-
* aipack-eval run [--suite <name>]... [--
|
|
7
|
-
* [--
|
|
8
|
-
* [--
|
|
6
|
+
* aipack-eval run [--suite <name>]... [--mode mock|live] [--model provider/id]
|
|
7
|
+
* [--api-key <key>] [--base-url <url>] [--temperature <0>]
|
|
8
|
+
* [--repeats <n>] [--cases-dir <dir>] [--report-dir <dir>]
|
|
9
|
+
* [--history <path>] [--baseline <path>] [--update-baseline]
|
|
10
|
+
* [--judge-model provider/id] [--embed-model <id>]
|
|
11
|
+
* aipack-eval compare --models m1,m2[,...] [--suite <name>]... [--mode live] ...
|
|
12
|
+
* aipack-eval import --file <export.json> [--out-dir <dir>] [--suite <name>]
|
|
13
|
+
* [--origin trace|bugfix] [--prefix <p>] [--dry-run]
|
|
14
|
+
* aipack-eval history [--path <file>] [--limit <n>] [--json]
|
|
9
15
|
*
|
|
10
16
|
* 退出码:0 = 全过且无回归;1 = 有失败或回归;2 = 用法错误。
|
|
11
17
|
*/
|
package/dist/src/cli.js
CHANGED
|
@@ -3,29 +3,79 @@
|
|
|
3
3
|
* packages/eval/src/cli.ts - aipack-eval 命令行入口
|
|
4
4
|
*
|
|
5
5
|
* 用法:
|
|
6
|
-
* aipack-eval run [--suite <name>]... [--
|
|
7
|
-
* [--
|
|
8
|
-
* [--
|
|
6
|
+
* aipack-eval run [--suite <name>]... [--mode mock|live] [--model provider/id]
|
|
7
|
+
* [--api-key <key>] [--base-url <url>] [--temperature <0>]
|
|
8
|
+
* [--repeats <n>] [--cases-dir <dir>] [--report-dir <dir>]
|
|
9
|
+
* [--history <path>] [--baseline <path>] [--update-baseline]
|
|
10
|
+
* [--judge-model provider/id] [--embed-model <id>]
|
|
11
|
+
* aipack-eval compare --models m1,m2[,...] [--suite <name>]... [--mode live] ...
|
|
12
|
+
* aipack-eval import --file <export.json> [--out-dir <dir>] [--suite <name>]
|
|
13
|
+
* [--origin trace|bugfix] [--prefix <p>] [--dry-run]
|
|
14
|
+
* aipack-eval history [--path <file>] [--limit <n>] [--json]
|
|
9
15
|
*
|
|
10
16
|
* 退出码:0 = 全过且无回归;1 = 有失败或回归;2 = 用法错误。
|
|
11
17
|
*/
|
|
12
18
|
import { loadCases } from './core/loader.js';
|
|
13
19
|
import { runEval } from './core/runner.js';
|
|
14
20
|
import { finalizeReport, renderMarkdown } from './core/report.js';
|
|
15
|
-
|
|
21
|
+
import { readHistory, renderHistoryTrend } from './core/history.js';
|
|
22
|
+
import { compareModels, renderComparisonMarkdown } from './core/compare.js';
|
|
23
|
+
import { importCases } from './core/import.js';
|
|
24
|
+
const USAGE = `用法: aipack-eval <run|compare|import|history> [options]
|
|
16
25
|
|
|
17
|
-
选项:
|
|
26
|
+
run 选项:
|
|
18
27
|
--suite <name> 只跑指定套件(可多次)
|
|
28
|
+
--mode <mock|live> 运行模式(缺省 mock;live = 真实 LLM)
|
|
29
|
+
--model <provider/id> live 模型,如 deepseek/deepseek-chat(或 AIPACK_EVAL_MODEL)
|
|
30
|
+
--api-key <key> live API Key(或 <PROVIDER>_API_KEY)
|
|
31
|
+
--base-url <url> live 端点覆盖(代理 / 兼容网关)
|
|
32
|
+
--temperature <x> live 采样温度(缺省 0)
|
|
33
|
+
--repeats <n> 每用例重复次数(live 缺省 3,pass@k 判定)
|
|
34
|
+
--concurrency <n> 并发用例数(缺省 8)
|
|
35
|
+
--timeout <ms> 单用例墙钟超时(缺省 30000)
|
|
36
|
+
--max-tokens <n> 全局 token 预算,超限熔断剩余用例
|
|
19
37
|
--cases-dir <dir> 用例目录(缺省 <包根>/eval/cases)
|
|
20
|
-
--report-dir <dir>
|
|
38
|
+
--report-dir <dir> 报告输出目录(缺省不写盘;'none' 不写盘)
|
|
39
|
+
--history <path> 历史趋势 JSONL(缺省 <report-dir>/history.jsonl)
|
|
21
40
|
--baseline <path> baseline 文件(门禁对比)
|
|
22
41
|
--update-baseline 用本次结果更新 baseline
|
|
23
42
|
--threshold <x> 通过率回归阈值(缺省 0.02)
|
|
24
|
-
--
|
|
43
|
+
--judge-model <p/id> LLM-as-judge 模型(须与被测模型异源;或 AIPACK_EVAL_JUDGE_MODEL)
|
|
44
|
+
--judge-api-key <key> judge API Key(或 AIPACK_EVAL_JUDGE_API_KEY)
|
|
45
|
+
--judge-base-url <url> judge 端点覆盖
|
|
46
|
+
--embed-model <id> semantic 评分器的 embedding 模型(或 AIPACK_EVAL_EMBEDDING_MODEL)
|
|
47
|
+
--embed-api-key <key> embedding API Key(或 AIPACK_EVAL_EMBEDDING_API_KEY / OPENAI_API_KEY)
|
|
48
|
+
--embed-base-url <url> embedding 端点覆盖(OpenAI 兼容 /embeddings)
|
|
49
|
+
|
|
50
|
+
compare 选项(L4 模型对比):
|
|
51
|
+
--models <m1,m2,...> 被测模型列表(逗号分隔,可多次;每模型跑同一组用例)
|
|
52
|
+
其余同 run(忽略 baseline / history)。
|
|
53
|
+
|
|
54
|
+
import 选项(trace 回流入库):
|
|
55
|
+
--file <path> export-eval 导出的 JSON(单 case / 数组 / {cases:[...]})
|
|
56
|
+
--out-dir <dir> 用例目录(缺省 <包根>/eval/cases)
|
|
57
|
+
--suite <name> 覆盖套件
|
|
58
|
+
--origin <o> 覆盖来源(trace | bugfix)
|
|
59
|
+
--prefix <p> case id 加前缀(如 bugfix)
|
|
60
|
+
--dry-run 只校验不写盘
|
|
61
|
+
|
|
62
|
+
history 选项:
|
|
63
|
+
--path <file> 历史文件(缺省 ./eval-results/history.jsonl)
|
|
64
|
+
--limit <n> 展示最近 N 次(缺省 15)
|
|
65
|
+
--json 输出原始 JSONL 条目
|
|
66
|
+
|
|
67
|
+
公共:
|
|
25
68
|
-h, --help 显示帮助`;
|
|
26
|
-
function
|
|
69
|
+
function num(raw, label) {
|
|
70
|
+
const v = Number(raw);
|
|
71
|
+
if (!Number.isFinite(v))
|
|
72
|
+
throw new Error(`${label} 必须为数字`);
|
|
73
|
+
return v;
|
|
74
|
+
}
|
|
75
|
+
function parseRunArgs(argv) {
|
|
27
76
|
const args = {
|
|
28
77
|
suites: [],
|
|
78
|
+
mode: 'mock',
|
|
29
79
|
updateBaseline: false,
|
|
30
80
|
threshold: 0.02,
|
|
31
81
|
};
|
|
@@ -40,18 +90,65 @@ function parseArgs(argv) {
|
|
|
40
90
|
};
|
|
41
91
|
try {
|
|
42
92
|
switch (a) {
|
|
43
|
-
case 'run':
|
|
44
|
-
args.command = 'run';
|
|
45
|
-
break;
|
|
46
93
|
case '--suite':
|
|
47
94
|
args.suites.push(next());
|
|
48
95
|
break;
|
|
96
|
+
case '--mode': {
|
|
97
|
+
const v = next();
|
|
98
|
+
if (v !== 'mock' && v !== 'live')
|
|
99
|
+
throw new Error('--mode 只能为 mock 或 live');
|
|
100
|
+
args.mode = v;
|
|
101
|
+
break;
|
|
102
|
+
}
|
|
103
|
+
case '--model':
|
|
104
|
+
args.model = next();
|
|
105
|
+
break;
|
|
106
|
+
case '--api-key':
|
|
107
|
+
args.apiKey = next();
|
|
108
|
+
break;
|
|
109
|
+
case '--base-url':
|
|
110
|
+
args.baseUrl = next();
|
|
111
|
+
break;
|
|
112
|
+
case '--temperature':
|
|
113
|
+
args.temperature = num(next(), '--temperature');
|
|
114
|
+
break;
|
|
115
|
+
case '--repeats': {
|
|
116
|
+
const v = Number(next());
|
|
117
|
+
if (!Number.isInteger(v) || v < 1)
|
|
118
|
+
throw new Error('repeats 须为正整数');
|
|
119
|
+
args.repeats = v;
|
|
120
|
+
break;
|
|
121
|
+
}
|
|
122
|
+
case '--concurrency': {
|
|
123
|
+
const v = Number(next());
|
|
124
|
+
if (!Number.isInteger(v) || v < 1)
|
|
125
|
+
throw new Error('concurrency 须为正整数');
|
|
126
|
+
args.concurrency = v;
|
|
127
|
+
break;
|
|
128
|
+
}
|
|
129
|
+
case '--timeout': {
|
|
130
|
+
const v = Number(next());
|
|
131
|
+
if (!Number.isFinite(v) || v <= 0)
|
|
132
|
+
throw new Error('timeout 须为正数(ms)');
|
|
133
|
+
args.timeoutMs = v;
|
|
134
|
+
break;
|
|
135
|
+
}
|
|
136
|
+
case '--max-tokens': {
|
|
137
|
+
const v = Number(next());
|
|
138
|
+
if (!Number.isFinite(v) || v <= 0)
|
|
139
|
+
throw new Error('max-tokens 须为正数');
|
|
140
|
+
args.maxTotalTokens = v;
|
|
141
|
+
break;
|
|
142
|
+
}
|
|
49
143
|
case '--cases-dir':
|
|
50
144
|
args.casesDir = next();
|
|
51
145
|
break;
|
|
52
146
|
case '--report-dir':
|
|
53
147
|
args.reportDir = next();
|
|
54
148
|
break;
|
|
149
|
+
case '--history':
|
|
150
|
+
args.historyPath = next();
|
|
151
|
+
break;
|
|
55
152
|
case '--baseline':
|
|
56
153
|
args.baselinePath = next();
|
|
57
154
|
break;
|
|
@@ -65,13 +162,30 @@ function parseArgs(argv) {
|
|
|
65
162
|
args.threshold = v;
|
|
66
163
|
break;
|
|
67
164
|
}
|
|
68
|
-
case '--
|
|
69
|
-
|
|
70
|
-
|
|
71
|
-
|
|
72
|
-
args.
|
|
165
|
+
case '--judge-model':
|
|
166
|
+
args.judgeModel = next();
|
|
167
|
+
break;
|
|
168
|
+
case '--judge-api-key':
|
|
169
|
+
args.judgeApiKey = next();
|
|
170
|
+
break;
|
|
171
|
+
case '--judge-base-url':
|
|
172
|
+
args.judgeBaseUrl = next();
|
|
173
|
+
break;
|
|
174
|
+
case '--embed-model':
|
|
175
|
+
args.embedModel = next();
|
|
176
|
+
break;
|
|
177
|
+
case '--embed-api-key':
|
|
178
|
+
args.embedApiKey = next();
|
|
179
|
+
break;
|
|
180
|
+
case '--embed-base-url':
|
|
181
|
+
args.embedBaseUrl = next();
|
|
182
|
+
break;
|
|
183
|
+
case '--models':
|
|
184
|
+
args.models = (args.models ?? []).concat(next()
|
|
185
|
+
.split(',')
|
|
186
|
+
.map((s) => s.trim())
|
|
187
|
+
.filter(Boolean));
|
|
73
188
|
break;
|
|
74
|
-
}
|
|
75
189
|
case '-h':
|
|
76
190
|
case '--help':
|
|
77
191
|
return { error: 'HELP' };
|
|
@@ -84,12 +198,10 @@ function parseArgs(argv) {
|
|
|
84
198
|
}
|
|
85
199
|
i += 1;
|
|
86
200
|
}
|
|
87
|
-
if (args.command !== 'run')
|
|
88
|
-
return { error: `缺少子命令 'run'\n${USAGE}` };
|
|
89
201
|
return args;
|
|
90
202
|
}
|
|
91
|
-
async function
|
|
92
|
-
const parsed =
|
|
203
|
+
async function cmdRun(argv) {
|
|
204
|
+
const parsed = parseRunArgs(argv);
|
|
93
205
|
if ('error' in parsed) {
|
|
94
206
|
if (parsed.error === 'HELP') {
|
|
95
207
|
console.log(USAGE);
|
|
@@ -109,22 +221,51 @@ async function main() {
|
|
|
109
221
|
console.error('没有可运行的用例');
|
|
110
222
|
return 1;
|
|
111
223
|
}
|
|
224
|
+
const reportDir = parsed.reportDir === 'none' ? undefined : parsed.reportDir;
|
|
225
|
+
const historyPath = parsed.historyPath ?? (reportDir ? `${reportDir.replace(/\/$/, '')}/history.jsonl` : undefined);
|
|
112
226
|
const config = {
|
|
113
|
-
mode:
|
|
227
|
+
mode: parsed.mode,
|
|
228
|
+
model: parsed.model,
|
|
229
|
+
apiKey: parsed.apiKey,
|
|
230
|
+
baseUrl: parsed.baseUrl,
|
|
231
|
+
temperature: parsed.temperature,
|
|
114
232
|
suites: parsed.suites,
|
|
115
233
|
repeats: parsed.repeats,
|
|
234
|
+
concurrency: parsed.concurrency,
|
|
235
|
+
timeoutMs: parsed.timeoutMs,
|
|
236
|
+
maxTotalTokens: parsed.maxTotalTokens,
|
|
116
237
|
casesDir: parsed.casesDir,
|
|
117
|
-
reportDir
|
|
238
|
+
reportDir,
|
|
239
|
+
historyPath,
|
|
118
240
|
baselinePath: parsed.baselinePath,
|
|
119
241
|
updateBaseline: parsed.updateBaseline,
|
|
120
242
|
regressionThreshold: parsed.threshold,
|
|
243
|
+
judgeModel: parsed.judgeModel,
|
|
244
|
+
judgeApiKey: parsed.judgeApiKey,
|
|
245
|
+
judgeBaseUrl: parsed.judgeBaseUrl,
|
|
246
|
+
embedModel: parsed.embedModel,
|
|
247
|
+
embedApiKey: parsed.embedApiKey,
|
|
248
|
+
embedBaseUrl: parsed.embedBaseUrl,
|
|
121
249
|
};
|
|
122
|
-
console.error(`加载 ${cases.length} 个用例${parsed.suites.length ? `(套件: ${parsed.suites.join(', ')})` : ''},mode
|
|
123
|
-
|
|
250
|
+
console.error(`加载 ${cases.length} 个用例${parsed.suites.length ? `(套件: ${parsed.suites.join(', ')})` : ''},mode=${parsed.mode}`);
|
|
251
|
+
let report;
|
|
252
|
+
try {
|
|
253
|
+
report = await runEval(cases, config);
|
|
254
|
+
}
|
|
255
|
+
catch (e) {
|
|
256
|
+
// live 装配失败(缺模型 / 缺 Key)属于使用错误
|
|
257
|
+
console.error(`live 模式配置错误: ${e.message}`);
|
|
258
|
+
return 2;
|
|
259
|
+
}
|
|
124
260
|
const finalized = await finalizeReport(report, config);
|
|
125
|
-
console.log(renderMarkdown(report));
|
|
261
|
+
console.log(renderMarkdown(report, { history: finalized.history }));
|
|
262
|
+
if (report.totals.cases === 0) {
|
|
263
|
+
console.error(`\n⚠️ 当前 mode=${parsed.mode} 下没有可运行的用例(${report.totals.skipped} 个被跳过),检查 --suite 与 --mode 是否匹配`);
|
|
264
|
+
}
|
|
126
265
|
if (finalized.jsonPath)
|
|
127
266
|
console.error(`\n报告已写入: ${finalized.jsonPath}`);
|
|
267
|
+
if (finalized.historyPath)
|
|
268
|
+
console.error(`历史已追加: ${finalized.historyPath}`);
|
|
128
269
|
if (config.updateBaseline && finalized.baseline === undefined) {
|
|
129
270
|
console.error(`baseline 已更新: ${config.baselinePath}`);
|
|
130
271
|
}
|
|
@@ -142,6 +283,200 @@ async function main() {
|
|
|
142
283
|
}
|
|
143
284
|
return exitCode;
|
|
144
285
|
}
|
|
286
|
+
async function cmdHistory(argv) {
|
|
287
|
+
let path = 'eval-results/history.jsonl';
|
|
288
|
+
let limit = 15;
|
|
289
|
+
let asJson = false;
|
|
290
|
+
for (let i = 0; i < argv.length; i++) {
|
|
291
|
+
const a = argv[i];
|
|
292
|
+
if (a === '--json') {
|
|
293
|
+
asJson = true;
|
|
294
|
+
}
|
|
295
|
+
else if (a === '--path') {
|
|
296
|
+
path = argv[++i] ?? path;
|
|
297
|
+
}
|
|
298
|
+
else if (a === '--limit') {
|
|
299
|
+
limit = Number(argv[++i] ?? limit);
|
|
300
|
+
}
|
|
301
|
+
else if (a === '-h' || a === '--help') {
|
|
302
|
+
console.log(USAGE);
|
|
303
|
+
return 0;
|
|
304
|
+
}
|
|
305
|
+
else {
|
|
306
|
+
console.error(`未知参数: ${a}\n${USAGE}`);
|
|
307
|
+
return 2;
|
|
308
|
+
}
|
|
309
|
+
}
|
|
310
|
+
const entries = await readHistory(path, limit);
|
|
311
|
+
if (entries.length === 0) {
|
|
312
|
+
console.error(`历史文件为空或不存在: ${path}`);
|
|
313
|
+
return 1;
|
|
314
|
+
}
|
|
315
|
+
if (asJson) {
|
|
316
|
+
console.log(entries.map((e) => JSON.stringify(e)).join('\n'));
|
|
317
|
+
return 0;
|
|
318
|
+
}
|
|
319
|
+
console.log(renderHistoryTrend(entries, limit));
|
|
320
|
+
return 0;
|
|
321
|
+
}
|
|
322
|
+
/** L4:同一组用例跨多模型对比(M5) */
|
|
323
|
+
async function cmdCompare(argv) {
|
|
324
|
+
const parsed = parseRunArgs(argv);
|
|
325
|
+
if ('error' in parsed) {
|
|
326
|
+
if (parsed.error === 'HELP') {
|
|
327
|
+
console.log(USAGE);
|
|
328
|
+
return 0;
|
|
329
|
+
}
|
|
330
|
+
console.error(parsed.error);
|
|
331
|
+
return 2;
|
|
332
|
+
}
|
|
333
|
+
const models = parsed.models ?? [];
|
|
334
|
+
if (models.length === 0) {
|
|
335
|
+
console.error(`compare 需要至少一个被测模型:--models deepseek/deepseek-chat,openai/gpt-4o-mini\n${USAGE}`);
|
|
336
|
+
return 2;
|
|
337
|
+
}
|
|
338
|
+
if (models.length !== new Set(models).size) {
|
|
339
|
+
console.error('compare 的 --models 含重复模型');
|
|
340
|
+
return 2;
|
|
341
|
+
}
|
|
342
|
+
const { cases, errors } = await loadCases(parsed.casesDir);
|
|
343
|
+
if (errors.length > 0) {
|
|
344
|
+
console.error('用例加载错误:');
|
|
345
|
+
for (const e of errors)
|
|
346
|
+
console.error(` - ${e}`);
|
|
347
|
+
return 1;
|
|
348
|
+
}
|
|
349
|
+
const config = {
|
|
350
|
+
mode: parsed.mode,
|
|
351
|
+
apiKey: parsed.apiKey,
|
|
352
|
+
baseUrl: parsed.baseUrl,
|
|
353
|
+
temperature: parsed.temperature,
|
|
354
|
+
suites: parsed.suites,
|
|
355
|
+
repeats: parsed.repeats,
|
|
356
|
+
concurrency: parsed.concurrency,
|
|
357
|
+
timeoutMs: parsed.timeoutMs,
|
|
358
|
+
maxTotalTokens: parsed.maxTotalTokens,
|
|
359
|
+
casesDir: parsed.casesDir,
|
|
360
|
+
judgeModel: parsed.judgeModel,
|
|
361
|
+
judgeApiKey: parsed.judgeApiKey,
|
|
362
|
+
judgeBaseUrl: parsed.judgeBaseUrl,
|
|
363
|
+
embedModel: parsed.embedModel,
|
|
364
|
+
embedApiKey: parsed.embedApiKey,
|
|
365
|
+
embedBaseUrl: parsed.embedBaseUrl,
|
|
366
|
+
};
|
|
367
|
+
console.error(`L4 对比:${models.length} 个模型 × ${cases.length} 个用例(mode=${parsed.mode}${parsed.suites.length ? `,套件: ${parsed.suites.join(', ')}` : ''})`);
|
|
368
|
+
let comparison;
|
|
369
|
+
try {
|
|
370
|
+
comparison = await compareModels(cases, config, models);
|
|
371
|
+
}
|
|
372
|
+
catch (e) {
|
|
373
|
+
console.error(`compare 配置错误: ${e.message}`);
|
|
374
|
+
return 2;
|
|
375
|
+
}
|
|
376
|
+
console.log(renderComparisonMarkdown(comparison));
|
|
377
|
+
// 报告写盘(compare-<ts>.md / .json)
|
|
378
|
+
if (parsed.reportDir && parsed.reportDir !== 'none') {
|
|
379
|
+
const { mkdir, writeFile } = await import('node:fs/promises');
|
|
380
|
+
const { join } = await import('node:path');
|
|
381
|
+
const dir = parsed.reportDir.replace(/\/$/, '');
|
|
382
|
+
const stamp = new Date().toISOString().replace(/[:.]/g, '-');
|
|
383
|
+
const mdPath = join(dir, `compare-${stamp}.md`);
|
|
384
|
+
const jsonPath = join(dir, `compare-${stamp}.json`);
|
|
385
|
+
await mkdir(dir, { recursive: true });
|
|
386
|
+
await writeFile(mdPath, renderComparisonMarkdown(comparison), 'utf-8');
|
|
387
|
+
await writeFile(jsonPath, JSON.stringify(comparison, null, 2), 'utf-8');
|
|
388
|
+
console.error(`\n对比报告已写入: ${mdPath}\n ${jsonPath}`);
|
|
389
|
+
}
|
|
390
|
+
return 0;
|
|
391
|
+
}
|
|
392
|
+
/** M4:trace 回流导出 JSON → eval/cases 入库 */
|
|
393
|
+
async function cmdImport(argv) {
|
|
394
|
+
let file;
|
|
395
|
+
let outDir;
|
|
396
|
+
let suite;
|
|
397
|
+
let origin;
|
|
398
|
+
let prefix;
|
|
399
|
+
let dryRun = false;
|
|
400
|
+
for (let i = 0; i < argv.length; i++) {
|
|
401
|
+
const a = argv[i];
|
|
402
|
+
const next = () => {
|
|
403
|
+
const v = argv[++i];
|
|
404
|
+
if (v === undefined)
|
|
405
|
+
throw new Error(`选项 ${a} 缺少参数值`);
|
|
406
|
+
return v;
|
|
407
|
+
};
|
|
408
|
+
try {
|
|
409
|
+
switch (a) {
|
|
410
|
+
case '--file':
|
|
411
|
+
file = next();
|
|
412
|
+
break;
|
|
413
|
+
case '--out-dir':
|
|
414
|
+
outDir = next();
|
|
415
|
+
break;
|
|
416
|
+
case '--suite':
|
|
417
|
+
suite = next();
|
|
418
|
+
break;
|
|
419
|
+
case '--origin': {
|
|
420
|
+
const v = next();
|
|
421
|
+
if (v !== 'trace' && v !== 'bugfix')
|
|
422
|
+
throw new Error('--origin 只能为 trace 或 bugfix');
|
|
423
|
+
origin = v;
|
|
424
|
+
break;
|
|
425
|
+
}
|
|
426
|
+
case '--prefix':
|
|
427
|
+
prefix = next();
|
|
428
|
+
break;
|
|
429
|
+
case '--dry-run':
|
|
430
|
+
dryRun = true;
|
|
431
|
+
break;
|
|
432
|
+
case '-h':
|
|
433
|
+
case '--help':
|
|
434
|
+
console.log(USAGE);
|
|
435
|
+
return 0;
|
|
436
|
+
default:
|
|
437
|
+
console.error(`未知参数: ${a}\n${USAGE}`);
|
|
438
|
+
return 2;
|
|
439
|
+
}
|
|
440
|
+
}
|
|
441
|
+
catch (e) {
|
|
442
|
+
console.error(`${e.message}\n${USAGE}`);
|
|
443
|
+
return 2;
|
|
444
|
+
}
|
|
445
|
+
}
|
|
446
|
+
if (!file) {
|
|
447
|
+
console.error(`import 需要 --file <export.json>\n${USAGE}`);
|
|
448
|
+
return 2;
|
|
449
|
+
}
|
|
450
|
+
const { written, errors } = await importCases({ file, outDir, suite, origin, prefix, dryRun });
|
|
451
|
+
if (errors.length > 0) {
|
|
452
|
+
console.error('入库错误:');
|
|
453
|
+
for (const e of errors)
|
|
454
|
+
console.error(` - ${e}`);
|
|
455
|
+
}
|
|
456
|
+
if (written.length > 0) {
|
|
457
|
+
console.error(dryRun ? '校验通过(dry-run,未写盘):' : `已入库 ${written.length} 个用例:`);
|
|
458
|
+
for (const w of written)
|
|
459
|
+
console.error(` - ${w}`);
|
|
460
|
+
}
|
|
461
|
+
return errors.length === 0 && written.length > 0 ? 0 : 1;
|
|
462
|
+
}
|
|
463
|
+
async function main() {
|
|
464
|
+
const [command, ...rest] = process.argv.slice(2);
|
|
465
|
+
if (command === undefined || command === '-h' || command === '--help') {
|
|
466
|
+
console.log(USAGE);
|
|
467
|
+
return command === undefined ? 2 : 0;
|
|
468
|
+
}
|
|
469
|
+
if (command === 'run')
|
|
470
|
+
return cmdRun(rest);
|
|
471
|
+
if (command === 'compare')
|
|
472
|
+
return cmdCompare(rest);
|
|
473
|
+
if (command === 'import')
|
|
474
|
+
return cmdImport(rest);
|
|
475
|
+
if (command === 'history')
|
|
476
|
+
return cmdHistory(rest);
|
|
477
|
+
console.error(`未知子命令: ${command}\n${USAGE}`);
|
|
478
|
+
return 2;
|
|
479
|
+
}
|
|
145
480
|
main().then((code) => process.exit(code), (e) => {
|
|
146
481
|
console.error(e);
|
|
147
482
|
process.exit(1);
|
package/dist/src/cli.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"cli.js","sourceRoot":"","sources":["../../src/cli.ts"],"names":[],"mappings":";AACA;;;;;;;;;GASG;AAEH,OAAO,EAAE,SAAS,EAAE,MAAM,eAAe,CAAC;AAC1C,OAAO,EAAE,OAAO,EAAE,MAAM,eAAe,CAAC;AACxC,OAAO,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AAc/D,MAAM,KAAK,GAAG;;;;;;;;;;8BAUgB,CAAC;AAE/B,SAAS,SAAS,CAAC,IAAc;IAC/B,MAAM,IAAI,GAAY;QACpB,MAAM,EAAE,EAAE;QACV,cAAc,EAAE,KAAK;QACrB,SAAS,EAAE,IAAI;KAChB,CAAC;IACF,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QACvB,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,MAAM,IAAI,GAAG,GAAW,EAAE;YACxB,MAAM,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;YACpB,IAAI,CAAC,KAAK,SAAS;gBAAE,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;YACtD,OAAO,CAAC,CAAC;QACX,CAAC,CAAC;QACF,IAAI,CAAC;YACH,QAAQ,CAAC,EAAE,CAAC;gBACV,KAAK,KAAK;oBACR,IAAI,CAAC,OAAO,GAAG,KAAK,CAAC;oBACrB,MAAM;gBACR,KAAK,SAAS;oBACZ,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,MAAM;gBACR,KAAK,aAAa;oBAChB,IAAI,CAAC,QAAQ,GAAG,IAAI,EAAE,CAAC;oBACvB,MAAM;gBACR,KAAK,cAAc;oBACjB,IAAI,CAAC,SAAS,GAAG,IAAI,EAAE,CAAC;oBACxB,MAAM;gBACR,KAAK,YAAY;oBACf,IAAI,CAAC,YAAY,GAAG,IAAI,EAAE,CAAC;oBAC3B,MAAM;gBACR,KAAK,mBAAmB;oBACtB,IAAI,CAAC,cAAc,GAAG,IAAI,CAAC;oBAC3B,MAAM;gBACR,KAAK,aAAa,CAAC,CAAC,CAAC;oBACnB,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC;wBAAE,MAAM,IAAI,KAAK,CAAC,UAAU,CAAC,CAAC;oBACvE,IAAI,CAAC,SAAS,GAAG,CAAC,CAAC;oBACnB,MAAM;gBACR,CAAC;gBACD,KAAK,WAAW,CAAC,CAAC,CAAC;oBACjB,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC;wBAAE,MAAM,IAAI,KAAK,CAAC,eAAe,CAAC,CAAC;oBACpE,IAAI,CAAC,OAAO,GAAG,CAAC,CAAC;oBACjB,MAAM;gBACR,CAAC;gBACD,KAAK,IAAI,CAAC;gBACV,KAAK,QAAQ;oBACX,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;gBAC3B;oBACE,OAAO,EAAE,KAAK,EAAE,SAAS,CAAC,KAAK,KAAK,EAAE,EAAE,CAAC;YAC7C,CAAC;QACH,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,OAAO,EAAE,KAAK,EAAE,GAAI,CAAW,CAAC,OAAO,KAAK,KAAK,EAAE,EAAE,CAAC;QACxD,CAAC;QACD,CAAC,IAAI,CAAC,CAAC;IACT,CAAC;IACD,IAAI,IAAI,CAAC,OAAO,KAAK,KAAK;QAAE,OAAO,EAAE,KAAK,EAAE,gBAAgB,KAAK,EAAE,EAAE,CAAC;IACtE,OAAO,IAAI,CAAC;AACd,CAAC;AAED,KAAK,UAAU,IAAI;IACjB,MAAM,MAAM,GAAG,SAAS,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAChD,IAAI,OAAO,IAAI,MAAM,EAAE,CAAC;QACtB,IAAI,MAAM,CAAC,KAAK,KAAK,MAAM,EAAE,CAAC;YAC5B,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACnB,OAAO,CAAC,CAAC;QACX,CAAC;QACD,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAC5B,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IAC3D,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;QACzB,KAAK,MAAM,CAAC,IAAI,MAAM;YAAE,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAClD,OAAO,CAAC,CAAC;IACX,CAAC;IACD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;QAC1B,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,MAAM,GAAc;QACxB,IAAI,EAAE,MAAM;QACZ,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,OAAO,EAAE,MAAM,CAAC,OAAO;QACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ;QACzB,SAAS,EAAE,MAAM,CAAC,SAAS,KAAK,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,SAAS;QACrE,YAAY,EAAE,MAAM,CAAC,YAAY;QACjC,cAAc,EAAE,MAAM,CAAC,cAAc;QACrC,mBAAmB,EAAE,MAAM,CAAC,SAAS;KACtC,CAAC;IAEF,OAAO,CAAC,KAAK,CACX,MAAM,KAAK,CAAC,MAAM,OAAO,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,YAAY,CACrG,CAAC;IACF,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IAC5C,MAAM,SAAS,GAAG,MAAM,cAAc,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAEvD,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC;IAEpC,IAAI,SAAS,CAAC,QAAQ;QAAE,OAAO,CAAC,KAAK,CAAC,YAAY,SAAS,CAAC,QAAQ,EAAE,CAAC,CAAC;IACxE,IAAI,MAAM,CAAC,cAAc,IAAI,SAAS,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;QAC9D,OAAO,CAAC,KAAK,CAAC,iBAAiB,MAAM,CAAC,YAAY,EAAE,CAAC,CAAC;IACxD,CAAC;IAED,IAAI,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,KAAK,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAEpE,IAAI,SAAS,CAAC,UAAU,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC,EAAE,EAAE,CAAC;QACrD,QAAQ,GAAG,CAAC,CAAC;QACb,MAAM,EAAE,YAAY,EAAE,WAAW,EAAE,GAAG,SAAS,CAAC,UAAU,CAAC;QAC3D,OAAO,CAAC,KAAK,CAAC,wBAAwB,YAAY,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,YAAY,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;QAC3G,KAAK,MAAM,CAAC,IAAI,WAAW,EAAE,CAAC;YAC5B,OAAO,CAAC,KAAK,CACX,OAAO,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,QAAQ,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CACvH,CAAC;QACJ,CAAC;IACH,CAAC;SAAM,IAAI,SAAS,CAAC,UAAU,EAAE,CAAC;QAChC,OAAO,CAAC,KAAK,CAAC,mBAAmB,CAAC,SAAS,CAAC,UAAU,CAAC,YAAY,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IAC7F,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,IAAI,EAAE,CAAC,IAAI,CACT,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAC5B,CAAC,CAAC,EAAE,EAAE;IACJ,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACjB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC,CACF,CAAC"}
|
|
1
|
+
{"version":3,"file":"cli.js","sourceRoot":"","sources":["../../src/cli.ts"],"names":[],"mappings":";AACA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,SAAS,EAAE,MAAM,eAAe,CAAC;AAC1C,OAAO,EAAE,OAAO,EAAE,MAAM,eAAe,CAAC;AACxC,OAAO,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,eAAe,CAAC;AAC/D,OAAO,EAAE,WAAW,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAC;AACjE,OAAO,EAAE,aAAa,EAAE,wBAAwB,EAAE,MAAM,gBAAgB,CAAC;AACzE,OAAO,EAAE,WAAW,EAAE,MAAM,eAAe,CAAC;AAiC5C,MAAM,KAAK,GAAG;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;8BA4CgB,CAAC;AAE/B,SAAS,GAAG,CAAC,GAAW,EAAE,KAAa;IACrC,MAAM,CAAC,GAAG,MAAM,CAAC,GAAG,CAAC,CAAC;IACtB,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,GAAG,KAAK,QAAQ,CAAC,CAAC;IAC3D,OAAO,CAAC,CAAC;AACX,CAAC;AAED,SAAS,YAAY,CAAC,IAAc;IAClC,MAAM,IAAI,GAAY;QACpB,MAAM,EAAE,EAAE;QACV,IAAI,EAAE,MAAM;QACZ,cAAc,EAAE,KAAK;QACrB,SAAS,EAAE,IAAI;KAChB,CAAC;IACF,IAAI,CAAC,GAAG,CAAC,CAAC;IACV,OAAO,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;QACvB,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,MAAM,IAAI,GAAG,GAAW,EAAE;YACxB,MAAM,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;YACpB,IAAI,CAAC,KAAK,SAAS;gBAAE,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;YACtD,OAAO,CAAC,CAAC;QACX,CAAC,CAAC;QACF,IAAI,CAAC;YACH,QAAQ,CAAC,EAAE,CAAC;gBACV,KAAK,SAAS;oBACZ,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,MAAM;gBACR,KAAK,QAAQ,CAAC,CAAC,CAAC;oBACd,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC;oBACjB,IAAI,CAAC,KAAK,MAAM,IAAI,CAAC,KAAK,MAAM;wBAAE,MAAM,IAAI,KAAK,CAAC,wBAAwB,CAAC,CAAC;oBAC5E,IAAI,CAAC,IAAI,GAAG,CAAC,CAAC;oBACd,MAAM;gBACR,CAAC;gBACD,KAAK,SAAS;oBACZ,IAAI,CAAC,KAAK,GAAG,IAAI,EAAE,CAAC;oBACpB,MAAM;gBACR,KAAK,WAAW;oBACd,IAAI,CAAC,MAAM,GAAG,IAAI,EAAE,CAAC;oBACrB,MAAM;gBACR,KAAK,YAAY;oBACf,IAAI,CAAC,OAAO,GAAG,IAAI,EAAE,CAAC;oBACtB,MAAM;gBACR,KAAK,eAAe;oBAClB,IAAI,CAAC,WAAW,GAAG,GAAG,CAAC,IAAI,EAAE,EAAE,eAAe,CAAC,CAAC;oBAChD,MAAM;gBACR,KAAK,WAAW,CAAC,CAAC,CAAC;oBACjB,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC;wBAAE,MAAM,IAAI,KAAK,CAAC,eAAe,CAAC,CAAC;oBACpE,IAAI,CAAC,OAAO,GAAG,CAAC,CAAC;oBACjB,MAAM;gBACR,CAAC;gBACD,KAAK,eAAe,CAAC,CAAC,CAAC;oBACrB,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,IAAI,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC;wBAAE,MAAM,IAAI,KAAK,CAAC,mBAAmB,CAAC,CAAC;oBACxE,IAAI,CAAC,WAAW,GAAG,CAAC,CAAC;oBACrB,MAAM;gBACR,CAAC;gBACD,KAAK,WAAW,CAAC,CAAC,CAAC;oBACjB,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;wBAAE,MAAM,IAAI,KAAK,CAAC,kBAAkB,CAAC,CAAC;oBACvE,IAAI,CAAC,SAAS,GAAG,CAAC,CAAC;oBACnB,MAAM;gBACR,CAAC;gBACD,KAAK,cAAc,CAAC,CAAC,CAAC;oBACpB,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;wBAAE,MAAM,IAAI,KAAK,CAAC,iBAAiB,CAAC,CAAC;oBACtE,IAAI,CAAC,cAAc,GAAG,CAAC,CAAC;oBACxB,MAAM;gBACR,CAAC;gBACD,KAAK,aAAa;oBAChB,IAAI,CAAC,QAAQ,GAAG,IAAI,EAAE,CAAC;oBACvB,MAAM;gBACR,KAAK,cAAc;oBACjB,IAAI,CAAC,SAAS,GAAG,IAAI,EAAE,CAAC;oBACxB,MAAM;gBACR,KAAK,WAAW;oBACd,IAAI,CAAC,WAAW,GAAG,IAAI,EAAE,CAAC;oBAC1B,MAAM;gBACR,KAAK,YAAY;oBACf,IAAI,CAAC,YAAY,GAAG,IAAI,EAAE,CAAC;oBAC3B,MAAM;gBACR,KAAK,mBAAmB;oBACtB,IAAI,CAAC,cAAc,GAAG,IAAI,CAAC;oBAC3B,MAAM;gBACR,KAAK,aAAa,CAAC,CAAC,CAAC;oBACnB,MAAM,CAAC,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;oBACzB,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC;wBAAE,MAAM,IAAI,KAAK,CAAC,UAAU,CAAC,CAAC;oBACvE,IAAI,CAAC,SAAS,GAAG,CAAC,CAAC;oBACnB,MAAM;gBACR,CAAC;gBACD,KAAK,eAAe;oBAClB,IAAI,CAAC,UAAU,GAAG,IAAI,EAAE,CAAC;oBACzB,MAAM;gBACR,KAAK,iBAAiB;oBACpB,IAAI,CAAC,WAAW,GAAG,IAAI,EAAE,CAAC;oBAC1B,MAAM;gBACR,KAAK,kBAAkB;oBACrB,IAAI,CAAC,YAAY,GAAG,IAAI,EAAE,CAAC;oBAC3B,MAAM;gBACR,KAAK,eAAe;oBAClB,IAAI,CAAC,UAAU,GAAG,IAAI,EAAE,CAAC;oBACzB,MAAM;gBACR,KAAK,iBAAiB;oBACpB,IAAI,CAAC,WAAW,GAAG,IAAI,EAAE,CAAC;oBAC1B,MAAM;gBACR,KAAK,kBAAkB;oBACrB,IAAI,CAAC,YAAY,GAAG,IAAI,EAAE,CAAC;oBAC3B,MAAM;gBACR,KAAK,UAAU;oBACb,IAAI,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,MAAM,IAAI,EAAE,CAAC,CAAC,MAAM,CACtC,IAAI,EAAE;yBACH,KAAK,CAAC,GAAG,CAAC;yBACV,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;yBACpB,MAAM,CAAC,OAAO,CAAC,CACnB,CAAC;oBACF,MAAM;gBACR,KAAK,IAAI,CAAC;gBACV,KAAK,QAAQ;oBACX,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;gBAC3B;oBACE,OAAO,EAAE,KAAK,EAAE,SAAS,CAAC,KAAK,KAAK,EAAE,EAAE,CAAC;YAC7C,CAAC;QACH,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,OAAO,EAAE,KAAK,EAAE,GAAI,CAAW,CAAC,OAAO,KAAK,KAAK,EAAE,EAAE,CAAC;QACxD,CAAC;QACD,CAAC,IAAI,CAAC,CAAC;IACT,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,KAAK,UAAU,MAAM,CAAC,IAAc;IAClC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAClC,IAAI,OAAO,IAAI,MAAM,EAAE,CAAC;QACtB,IAAI,MAAM,CAAC,KAAK,KAAK,MAAM,EAAE,CAAC;YAC5B,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACnB,OAAO,CAAC,CAAC;QACX,CAAC;QACD,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAC5B,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IAC3D,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;QACzB,KAAK,MAAM,CAAC,IAAI,MAAM;YAAE,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAClD,OAAO,CAAC,CAAC;IACX,CAAC;IACD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACvB,OAAO,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;QAC1B,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,SAAS,GAAG,MAAM,CAAC,SAAS,KAAK,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC;IAC7E,MAAM,WAAW,GACf,MAAM,CAAC,WAAW,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,GAAG,SAAS,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,gBAAgB,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC;IAElG,MAAM,MAAM,GAAc;QACxB,IAAI,EAAE,MAAM,CAAC,IAAI;QACjB,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,OAAO,EAAE,MAAM,CAAC,OAAO;QACvB,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,OAAO,EAAE,MAAM,CAAC,OAAO;QACvB,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,cAAc,EAAE,MAAM,CAAC,cAAc;QACrC,QAAQ,EAAE,MAAM,CAAC,QAAQ;QACzB,SAAS;QACT,WAAW;QACX,YAAY,EAAE,MAAM,CAAC,YAAY;QACjC,cAAc,EAAE,MAAM,CAAC,cAAc;QACrC,mBAAmB,EAAE,MAAM,CAAC,SAAS;QACrC,UAAU,EAAE,MAAM,CAAC,UAAU;QAC7B,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,YAAY,EAAE,MAAM,CAAC,YAAY;QACjC,UAAU,EAAE,MAAM,CAAC,UAAU;QAC7B,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,YAAY,EAAE,MAAM,CAAC,YAAY;KAClC,CAAC;IAEF,OAAO,CAAC,KAAK,CACX,MAAM,KAAK,CAAC,MAAM,OAAO,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,SAAS,MAAM,CAAC,IAAI,EAAE,CAC/G,CAAC;IAEF,IAAI,MAAkB,CAAC;IACvB,IAAI,CAAC;QACH,MAAM,GAAG,MAAM,OAAO,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IACxC,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,+BAA+B;QAC/B,OAAO,CAAC,KAAK,CAAC,gBAAiB,CAAW,CAAC,OAAO,EAAE,CAAC,CAAC;QACtD,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,SAAS,GAAG,MAAM,cAAc,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IACvD,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC,MAAM,EAAE,EAAE,OAAO,EAAE,SAAS,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;IAEpE,IAAI,MAAM,CAAC,MAAM,CAAC,KAAK,KAAK,CAAC,EAAE,CAAC;QAC9B,OAAO,CAAC,KAAK,CACX,iBAAiB,MAAM,CAAC,IAAI,cAAc,MAAM,CAAC,MAAM,CAAC,OAAO,iCAAiC,CACjG,CAAC;IACJ,CAAC;IAED,IAAI,SAAS,CAAC,QAAQ;QAAE,OAAO,CAAC,KAAK,CAAC,YAAY,SAAS,CAAC,QAAQ,EAAE,CAAC,CAAC;IACxE,IAAI,SAAS,CAAC,WAAW;QAAE,OAAO,CAAC,KAAK,CAAC,UAAU,SAAS,CAAC,WAAW,EAAE,CAAC,CAAC;IAC5E,IAAI,MAAM,CAAC,cAAc,IAAI,SAAS,CAAC,QAAQ,KAAK,SAAS,EAAE,CAAC;QAC9D,OAAO,CAAC,KAAK,CAAC,iBAAiB,MAAM,CAAC,YAAY,EAAE,CAAC,CAAC;IACxD,CAAC;IAED,IAAI,QAAQ,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,KAAK,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAEpE,IAAI,SAAS,CAAC,UAAU,IAAI,CAAC,SAAS,CAAC,UAAU,CAAC,EAAE,EAAE,CAAC;QACrD,QAAQ,GAAG,CAAC,CAAC;QACb,MAAM,EAAE,YAAY,EAAE,WAAW,EAAE,GAAG,SAAS,CAAC,UAAU,CAAC;QAC3D,OAAO,CAAC,KAAK,CACX,wBAAwB,YAAY,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,YAAY,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,CAC5F,CAAC;QACF,KAAK,MAAM,CAAC,IAAI,WAAW,EAAE,CAAC;YAC5B,OAAO,CAAC,KAAK,CACX,OAAO,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,CAAC,QAAQ,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CACvH,CAAC;QACJ,CAAC;IACH,CAAC;SAAM,IAAI,SAAS,CAAC,UAAU,EAAE,CAAC;QAChC,OAAO,CAAC,KAAK,CAAC,mBAAmB,CAAC,SAAS,CAAC,UAAU,CAAC,YAAY,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;IAC7F,CAAC;IAED,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,KAAK,UAAU,UAAU,CAAC,IAAc;IACtC,IAAI,IAAI,GAAG,4BAA4B,CAAC;IACxC,IAAI,KAAK,GAAG,EAAE,CAAC;IACf,IAAI,MAAM,GAAG,KAAK,CAAC;IACnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC;YACnB,MAAM,GAAG,IAAI,CAAC;QAChB,CAAC;aAAM,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC;YAC1B,IAAI,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,IAAI,CAAC;QAC3B,CAAC;aAAM,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAC3B,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC;QACrC,CAAC;aAAM,IAAI,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC;YACxC,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACnB,OAAO,CAAC,CAAC;QACX,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,KAAK,EAAE,CAAC,CAAC;YACtC,OAAO,CAAC,CAAC;QACX,CAAC;IACH,CAAC;IAED,MAAM,OAAO,GAAG,MAAM,WAAW,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;IAC/C,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,CAAC,KAAK,CAAC,eAAe,IAAI,EAAE,CAAC,CAAC;QACrC,OAAO,CAAC,CAAC;IACX,CAAC;IACD,IAAI,MAAM,EAAE,CAAC;QACX,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;QAC9D,OAAO,CAAC,CAAC;IACX,CAAC;IACD,OAAO,CAAC,GAAG,CAAC,kBAAkB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC,CAAC;IAChD,OAAO,CAAC,CAAC;AACX,CAAC;AAED,yBAAyB;AACzB,KAAK,UAAU,UAAU,CAAC,IAAc;IACtC,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAClC,IAAI,OAAO,IAAI,MAAM,EAAE,CAAC;QACtB,IAAI,MAAM,CAAC,KAAK,KAAK,MAAM,EAAE,CAAC;YAC5B,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;YACnB,OAAO,CAAC,CAAC;QACX,CAAC;QACD,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QAC5B,OAAO,CAAC,CAAC;IACX,CAAC;IACD,MAAM,MAAM,GAAG,MAAM,CAAC,MAAM,IAAI,EAAE,CAAC;IACnC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,CAAC,0EAA0E,KAAK,EAAE,CAAC,CAAC;QACjG,OAAO,CAAC,CAAC;IACX,CAAC;IACD,IAAI,MAAM,CAAC,MAAM,KAAK,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;QAC3C,OAAO,CAAC,KAAK,CAAC,0BAA0B,CAAC,CAAC;QAC1C,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,MAAM,SAAS,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IAC3D,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC;QACzB,KAAK,MAAM,CAAC,IAAI,MAAM;YAAE,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;QAClD,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,MAAM,GAAc;QACxB,IAAI,EAAE,MAAM,CAAC,IAAI;QACjB,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,OAAO,EAAE,MAAM,CAAC,OAAO;QACvB,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,OAAO,EAAE,MAAM,CAAC,OAAO;QACvB,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,SAAS,EAAE,MAAM,CAAC,SAAS;QAC3B,cAAc,EAAE,MAAM,CAAC,cAAc;QACrC,QAAQ,EAAE,MAAM,CAAC,QAAQ;QACzB,UAAU,EAAE,MAAM,CAAC,UAAU;QAC7B,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,YAAY,EAAE,MAAM,CAAC,YAAY;QACjC,UAAU,EAAE,MAAM,CAAC,UAAU;QAC7B,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,YAAY,EAAE,MAAM,CAAC,YAAY;KAClC,CAAC;IAEF,OAAO,CAAC,KAAK,CACX,SAAS,MAAM,CAAC,MAAM,UAAU,KAAK,CAAC,MAAM,aAAa,MAAM,CAAC,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,CACzI,CAAC;IAEF,IAAI,UAAU,CAAC;IACf,IAAI,CAAC;QACH,UAAU,GAAG,MAAM,aAAa,CAAC,KAAK,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC;IAC1D,CAAC;IAAC,OAAO,CAAC,EAAE,CAAC;QACX,OAAO,CAAC,KAAK,CAAC,iBAAkB,CAAW,CAAC,OAAO,EAAE,CAAC,CAAC;QACvD,OAAO,CAAC,CAAC;IACX,CAAC;IAED,OAAO,CAAC,GAAG,CAAC,wBAAwB,CAAC,UAAU,CAAC,CAAC,CAAC;IAElD,gCAAgC;IAChC,IAAI,MAAM,CAAC,SAAS,IAAI,MAAM,CAAC,SAAS,KAAK,MAAM,EAAE,CAAC;QACpD,MAAM,EAAE,KAAK,EAAE,SAAS,EAAE,GAAG,MAAM,MAAM,CAAC,kBAAkB,CAAC,CAAC;QAC9D,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,MAAM,CAAC,WAAW,CAAC,CAAC;QAC3C,MAAM,GAAG,GAAG,MAAM,CAAC,SAAS,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAChD,MAAM,KAAK,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC;QAC7D,MAAM,MAAM,GAAG,IAAI,CAAC,GAAG,EAAE,WAAW,KAAK,KAAK,CAAC,CAAC;QAChD,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,WAAW,KAAK,OAAO,CAAC,CAAC;QACpD,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QACtC,MAAM,SAAS,CAAC,MAAM,EAAE,wBAAwB,CAAC,UAAU,CAAC,EAAE,OAAO,CAAC,CAAC;QACvE,MAAM,SAAS,CAAC,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,UAAU,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;QACxE,OAAO,CAAC,KAAK,CAAC,cAAc,MAAM,qBAAqB,QAAQ,EAAE,CAAC,CAAC;IACrE,CAAC;IAED,OAAO,CAAC,CAAC;AACX,CAAC;AAED,yCAAyC;AACzC,KAAK,UAAU,SAAS,CAAC,IAAc;IACrC,IAAI,IAAwB,CAAC;IAC7B,IAAI,MAA0B,CAAC;IAC/B,IAAI,KAAyB,CAAC;IAC9B,IAAI,MAA8B,CAAC;IACnC,IAAI,MAA0B,CAAC;IAC/B,IAAI,MAAM,GAAG,KAAK,CAAC;IAEnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QAClB,MAAM,IAAI,GAAG,GAAW,EAAE;YACxB,MAAM,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;YACpB,IAAI,CAAC,KAAK,SAAS;gBAAE,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;YACtD,OAAO,CAAC,CAAC;QACX,CAAC,CAAC;QACF,IAAI,CAAC;YACH,QAAQ,CAAC,EAAE,CAAC;gBACV,KAAK,QAAQ;oBACX,IAAI,GAAG,IAAI,EAAE,CAAC;oBACd,MAAM;gBACR,KAAK,WAAW;oBACd,MAAM,GAAG,IAAI,EAAE,CAAC;oBAChB,MAAM;gBACR,KAAK,SAAS;oBACZ,KAAK,GAAG,IAAI,EAAE,CAAC;oBACf,MAAM;gBACR,KAAK,UAAU,CAAC,CAAC,CAAC;oBAChB,MAAM,CAAC,GAAG,IAAI,EAAE,CAAC;oBACjB,IAAI,CAAC,KAAK,OAAO,IAAI,CAAC,KAAK,QAAQ;wBAAE,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAC;oBACpF,MAAM,GAAG,CAAC,CAAC;oBACX,MAAM;gBACR,CAAC;gBACD,KAAK,UAAU;oBACb,MAAM,GAAG,IAAI,EAAE,CAAC;oBAChB,MAAM;gBACR,KAAK,WAAW;oBACd,MAAM,GAAG,IAAI,CAAC;oBACd,MAAM;gBACR,KAAK,IAAI,CAAC;gBACV,KAAK,QAAQ;oBACX,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;oBACnB,OAAO,CAAC,CAAC;gBACX;oBACE,OAAO,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,KAAK,EAAE,CAAC,CAAC;oBACtC,OAAO,CAAC,CAAC;YACb,CAAC;QACH,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,OAAO,CAAC,KAAK,CAAC,GAAI,CAAW,CAAC,OAAO,KAAK,KAAK,EAAE,CAAC,CAAC;YACnD,OAAO,CAAC,CAAC;QACX,CAAC;IACH,CAAC;IAED,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,OAAO,CAAC,KAAK,CAAC,mCAAmC,KAAK,EAAE,CAAC,CAAC;QAC1D,OAAO,CAAC,CAAC;IACX,CAAC;IAED,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,GAAG,MAAM,WAAW,CAAC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC;IAC/F,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACtB,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QACvB,KAAK,MAAM,CAAC,IAAI,MAAM;YAAE,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACpD,CAAC;IACD,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACvB,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,oBAAoB,CAAC,CAAC,CAAC,OAAO,OAAO,CAAC,MAAM,OAAO,CAAC,CAAC;QAC5E,KAAK,MAAM,CAAC,IAAI,OAAO;YAAE,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;IACrD,CAAC;IACD,OAAO,MAAM,CAAC,MAAM,KAAK,CAAC,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC3D,CAAC;AAED,KAAK,UAAU,IAAI;IACjB,MAAM,CAAC,OAAO,EAAE,GAAG,IAAI,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACjD,IAAI,OAAO,KAAK,SAAS,IAAI,OAAO,KAAK,IAAI,IAAI,OAAO,KAAK,QAAQ,EAAE,CAAC;QACtE,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC;QACnB,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACvC,CAAC;IACD,IAAI,OAAO,KAAK,KAAK;QAAE,OAAO,MAAM,CAAC,IAAI,CAAC,CAAC;IAC3C,IAAI,OAAO,KAAK,SAAS;QAAE,OAAO,UAAU,CAAC,IAAI,CAAC,CAAC;IACnD,IAAI,OAAO,KAAK,QAAQ;QAAE,OAAO,SAAS,CAAC,IAAI,CAAC,CAAC;IACjD,IAAI,OAAO,KAAK,SAAS;QAAE,OAAO,UAAU,CAAC,IAAI,CAAC,CAAC;IACnD,OAAO,CAAC,KAAK,CAAC,UAAU,OAAO,KAAK,KAAK,EAAE,CAAC,CAAC;IAC7C,OAAO,CAAC,CAAC;AACX,CAAC;AAED,IAAI,EAAE,CAAC,IAAI,CACT,CAAC,IAAI,EAAE,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAC5B,CAAC,CAAC,EAAE,EAAE;IACJ,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACjB,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAClB,CAAC,CACF,CAAC"}
|
|
@@ -0,0 +1,28 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* packages/eval/src/core/compare.ts - L4 模型 / Provider 对比(M5)
|
|
3
|
+
*
|
|
4
|
+
* 同一组用例在多个被测模型上各跑一遍 runEval(顺序执行,各算各的
|
|
5
|
+
* 预算与 repeats),汇总出通过率 / 平均分 / token 成本 / 延迟 /
|
|
6
|
+
* 分套件通过率矩阵。judge / embed 装配沿用 base config(judge 异源
|
|
7
|
+
* 校验对每个被测模型独立生效)。
|
|
8
|
+
*/
|
|
9
|
+
import type { EvalCase, EvalReport, RunConfig } from './types';
|
|
10
|
+
export interface ModelComparisonEntry {
|
|
11
|
+
/** 被测模型标识('provider/modelId') */
|
|
12
|
+
model: string;
|
|
13
|
+
report: EvalReport;
|
|
14
|
+
}
|
|
15
|
+
export interface ComparisonReport {
|
|
16
|
+
startedAt: string;
|
|
17
|
+
durationMs: number;
|
|
18
|
+
mode: 'mock' | 'live';
|
|
19
|
+
suites: string[];
|
|
20
|
+
entries: ModelComparisonEntry[];
|
|
21
|
+
}
|
|
22
|
+
/** 逐个模型跑同一组用例(顺序执行,避免跨模型并发互扰限流) */
|
|
23
|
+
export declare function compareModels(cases: EvalCase[], base: RunConfig, models: string[]): Promise<ComparisonReport>;
|
|
24
|
+
/**
|
|
25
|
+
* 对比表(报告按被测模型逐列对比;passRate 用 runEval 同一口径:
|
|
26
|
+
* 跳过用例不计入分母)。
|
|
27
|
+
*/
|
|
28
|
+
export declare function renderComparisonMarkdown(cmp: ComparisonReport): string;
|